Re: Typesetting Lojban [was: Lojban word processor for Windows?]
xod
> Do you like the fact that all Lojban gismu are of equal length,
> measured in phonemes/letters (5 in all, 3 consonants and 2 vowels)?
> Well, I want the next logical thing: I also want them to be of
> equal width on paper. That doesn't have to mean monospace as in
> all characters being of the same width, but I'd like every consonant
> to be as wide as every other, and every vowel as wide as every other
> (perhaps less wide than the consonants). It makes sense for {' , .}
> to be much narrower than the letters.
Of course, with these conditions the font is no longer monospace.
Check out Lojban set in Caslon, though.
-----
Perpetual Progress, Self-Transformation, Practical Optimism, Intelligent Technology,
Open Society, Self-Direction, and Rational Thinking.
David Brookshire Conner
> brook writes:
> >Most of the basic little words end up being a single glyph: one tengwa
> >with one tehta above, maybe one below. This suggests ligatures for
> >these combinations, at least in Tengwar. They become pictogram-like
> >word-characters, yet still have all the phonemic information apparent.
>
> You really should have come up with some way to make diphthongs -- both VV
> and V'V -- into tehta-combinations. That way *all* cmavo would be single
> symbols. It would make a nice cmavo/brivla distinction. Not sure if it's
> possible without making a real mess of the notation, though.
Ah, didn't mention that detail, but had in fact thought of it.
In my tengwar mode for lojban, the tehta above a tengwar is for a
following vowel, as most lojban words end in a vowel. As in Tolkien's
own Tengwar mode for Old English, a second following vowel is
represented by a tehta *under* the tengwa. Because of the shape of
the tehtar, you can either draw the lower tehta right-side up, or
upside down (i.e., mirrored, which appeals to me esthetically).
So that covers cvv. ' is a tengwa itself (you should like that, Mark -
really emphasizes it) - it's a voiceless velar fricative (x is the
voiced velar fricative). So that handles cv'v with two tengwar and two
tehta - still visually clear from a selbri, which *must* have a bare
tengwa somewhere. v or vv cmavo can be represented using the short
carrier symbol.
So what about v'v? You could either write this as a short carrier
followed by the ' tengwa or (if no ambiguity resulted) you could break
the following-vowel pattern and put the first v above the ' tengwa and
the second one underneath (since ' never comes between a break and a
vowel or between a consonant and a vowel, this works more than might
first be apparent).
> >Actually, the one open question I'd still muddling about for this mode
> >is whether to put "r" and "l" in Row 6 of the Tengwar - the
> >semivowels. This has a certain elegance, but it does make the writing
> >a bit more monotonous looking. But most tengwar modes don't put r and
> >l up there, but as part of the "other tengwa".
>
> Tengwar by their nature are monotonous-looking when written.
True. Though Quenya uses enough of the "other" tengwar that it looks
less monotonous than lojban does.
> Probably the
> weakest thing, aesthetically and practically, about the system: most of the
> letters resemble one another. You can't get around that, might as well
> live with it.
>
> There's a certain charm to teeny little Grade 6 r/l, especially given their
> role as hyphens/glue in Lojban.
Actually, the hyphen/glue thing makes me think they *shouldn't* be
grade 6 - if they look visibly different, they suggest a different
structure as well. In fact, this is one place where a Loglan mode
works a little better - you can use Grade 6 for R and L when it's
acting like a vowel, and the other tengwar otherwise.
> Have you bounced this off Ivan Derzhanski, Lojbanist, typographer, and
> tengwarist extraordinaire?
Nope. Don't know him. Got an email address? Or is he on this list?
> >I guess this seems ligature-heavy, but to my thinking, typing the
> >letters together emphasizes the structure, which is part of the beauty
> >of lojban.
>
> Come to think of it, you may be on to something here. Using ligatures for
> consonant-clusters emphasizes the clustering, and makes them distinctive.
Yeah - just like the bare tengwa does when using tengwar.
> And noticing consonant clusters is critical in Lojban (especially as
> written with frequent "compound cmavo"). And maybe that's what you just
> said.
Thanks - it is! :-)
> Ligating cmavo might dilute that, but not if done carefully.
It might, but I can think of two approaches to doing it carefully.
One is to somehow style the consonant-vowel ligatures for cmavo
differently. Maybe the i looks more like a slash, or some other subtle
tweak to the serifs or shapes.
The other is to simply not use ligatures with consonants in cmavo -
there's plenty you can use that are just vowels (or vowels and ').
> Not much ligating you can do with {.i} though. Except for a special font
> or form for the {i}.
Which is all a ligature is anyway :-) An i with a stem that looks
something like a backwards capitol L.
> Bear in mind I'm one of those who pushed early on for permitting {h} as an
> alloglyph of {'}, for handwriting. So I'm in favor of pretty heavy
> representations of {'}, and generally fearful of losing it. Mmm, I'm torn
> regarding the idea of merging the apostrophe and the dot. If anything, I
> want to emphasize the break, not obscure it. But on the other hand, I
> could see some kind of big bold comma-ish thing ligated on top of an {i}
> that almost becomes a syllable-glyph for {'i} (the /hi/ syllable).
Exactly.
> Confusion with {ta' i} or {ta i} doesn't scare me much, since I do NOT
> intend for the {'} to be lost (if anything I would overemphasize it), and
> if we're looking to pump up the visual distinctiveness of the {.i} cmavo it
> will not conflict with the {i} in {ta'i}. All of which also doesn't treat
> the other four vowels and the need for a certain amount of visual
> consistency.
'a is easy, as is 'u. 'o and 'e are a bit more of a stretch, but if
you think about the bar on a Q or the hook on double-loop forms of g,
you get something pretty reasonable. And they all involve this big,
comma-like hook thing.
Ligating i with a preceding vowel is an easy one, as is u and a (if
you use the small-loop-overhook shape of a, not the single loop
form). e and o are harder, but again, not impossible. For example, eo
ligates much like ae. ao is an obvious extension. And so on....
> Don't fear heavy ligatures; wait till I show you my Klingon font...
:-)
Brook
---------
Hidden DOS secret: add BUGS=OFF to your CONFIG.SYS
---------
Fancy. Myth. Magic.
http://www.concentric.net/~nellardo/
ma...@kli.org
>Date: Sat, 18 Sep 1999 20:06:58 -0400 (EDT)
>Cc: loj...@onelist.com
>
>From: David Brookshire Conner <nell...@concentric.net>
>
>ma...@kli.org writes:
> > brook writes:
> > >Most of the basic little words end up being a single glyph: one tengwa
> > >with one tehta above, maybe one below. This suggests ligatures for
> > >these combinations, at least in Tengwar. They become pictogram-like
> > >word-characters, yet still have all the phonemic information apparent.
> >
> > You really should have come up with some way to make diphthongs -- both VV
> > and V'V -- into tehta-combinations. That way *all* cmavo would be single
> > symbols. It would make a nice cmavo/brivla distinction. Not sure if it's
> > possible without making a real mess of the notation, though.
>So that covers cvv. ' is a tengwa itself (you should like that, Mark -
>really emphasizes it) - it's a voiceless velar fricative (x is the
>voiced velar fricative). So that handles cv'v with two tengwar and two
>tehta - still visually clear from a selbri, which *must* have a bare
>tengwa somewhere. v or vv cmavo can be represented using the short
>carrier symbol.
I like the emphasis... but considering how much all tengwar look alike it
may be overdoing it; I might be with Ivan here. Actually, halla would
probably be a fine choice for {'}; isn't that what's suggested in The
Book's tengwar mode? Or else hyarmen, for distinctiveness.
> > >Actually, the one open question I'd still muddling about for this mode
> > >is whether to put "r" and "l" in Row 6 of the Tengwar - the
> > >semivowels. This has a certain elegance, but it does make the writing
> > >a bit more monotonous looking. But most tengwar modes don't put r and
> > >l up there, but as part of the "other tengwa".
> >
> > Tengwar by their nature are monotonous-looking when written.
>
>True. Though Quenya uses enough of the "other" tengwar that it looks
>less monotonous than lojban does.
Monotony isn't the worst of it. Even if there are plenty of "extras", you
still have the problem that two very different words look almost identical
due to the similar tengwar. Hebrew gets criticized for this for its few
pairs of similar letters; it's far worse in Quenya.
> > There's a certain charm to teeny little Grade 6 r/l, especially given their
> > role as hyphens/glue in Lojban.
>
>Actually, the hyphen/glue thing makes me think they *shouldn't* be
>grade 6 - if they look visibly different, they suggest a different
>structure as well. In fact, this is one place where a Loglan mode
>works a little better - you can use Grade 6 for R and L when it's
>acting like a vowel, and the other tengwar otherwise.
But the grade-6, as I recall, are small little things, good for hyphens.
The other letters are more consonantal.
Oh, and {y} should definitely be somehow distinctive, in any mode (even
English works this OK, with the descender): it signals "brivla" just like a
consonant cluster. Oh, and I was thinking about your reliance on
tehta-less tengwar... that's okay, but bear in mind that noticing the
*absence* of a tehta is a lot harder than noticing the *presence* of
something. It works well at the ends of words for flagging cmene, but may
be easier to miss medially.
> > Have you bounced this off Ivan Derzhanski, Lojbanist, typographer, and
> > tengwarist extraordinaire?
>
>Nope. Don't know him. Got an email address? Or is he on this list?
Don't find him, he'll find you.
> > Not much ligating you can do with {.i} though. Except for a special font
> > or form for the {i}.
>
>Which is all a ligature is anyway :-) An i with a stem that looks
>something like a backwards capitol L.
Or larger, or bold, or swash...
~mark
Robert J. Chassell
... studies on font readability (as I recall), indicate that what
you find easiest to read is what you learn to read on. ...
On the contrary, studies I read some years ago suggested that while
practice is important, given equivalent degrees of practice, the
information content of the interface makes the difference.
If it is easier for reader to detect and recognize a difference among
letters or words, it is easier to read. Hence, given equal experience
among readers, fonts with serifs were easier, since they provides more
information per character and per word.
--
Robert J. Chassell b...@rattlesnake.com
Rattlesnake Enterprises http://www.rattlesnake.com
Bob LeChevalier (lojbab)
>From: "Robert J. Chassell" <b...@rattlesnake.com>
>David Brookshire Conner <nell...@concentric.net> wrote:
>
> ... studies on font readability (as I recall), indicate that what
> you find easiest to read is what you learn to read on. ...
>
>On the contrary, studies I read some years ago suggested that while
>practice is important, given equivalent degrees of practice, the
>information content of the interface makes the difference.
>
>If it is easier for reader to detect and recognize a difference among
>letters or words, it is easier to read. Hence, given equal experience
>among readers, fonts with serifs were easier, since they provides more
>information per character and per word.
In particular, there has been a historical problem with Loglan/Lojban text
that is sans serif, because there is typically too little difference
between sans lower case ell and sans upper case eye (l and I). TLI Loglan
capitalizes the first word of every sentence, and the Loglan 1 book has
many examples where there could be confusion between the words "la" and
"Ia", as well as "le" and "Ie", found at the start of sentences.
lojbab
----
lojbab ***NOTE NEW ADDRESS*** loj...@lojban.org
Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273
Artificial language Loglan/Lojban:
see Lojban WWW Server: href=" http://xiron.pc.helsinki.fi/lojban/ "
Order _The Complete Lojban Language_ - see our Web pages or ask me.
Ivan A Derzhanski
(1) Like Mark, I prefer monospace to proportional for Lojban.
I don't know whether Lojban is hot or cold, poetic or prosaic,
fluid or solid, nor whether it needs to be any of those things.
Its first and strongest claim is to being a *logical* language,
and in this it can't afford to fail.
Forget about typewritten English; think on a larger scale.
The Chinese script is always monospace, and yet looks gorgeous,
doesn't it? More to the point, it is a logical thing: 1 unit
of width = 1 syllable = 1 morpheme. (With very few exceptions.)
Do you like the fact that all Lojban gismu are of equal length,
measured in phonemes/letters (5 in all, 3 consonants and 2 vowels)?
Well, I want the next logical thing: I also want them to be of
equal width on paper. That doesn't have to mean monospace as in
all characters being of the same width, but I'd like every consonant
to be as wide as every other, and every vowel as wide as every other
(perhaps less wide than the consonants). It makes sense for {' , .}
to be much narrower than the letters.
(2) _cm_ does occur in English, but is very rare; try a search
for `*cm*' in <http://www.m-w.com/cgi-bin/mweb>. I don't mind
ligatures in general, but I don't think I'd want any in Lojban
-- for the same Vulcanish reason as above: the distinction
that would be introduced between ligatured and unligatured
sequences would be illogical and artificial, and by virtue
of that fact most undesirable in the context of Lojban.
(3) Lojban in tengwar should work, although, in fairness, there are
natlangs whose consonant structures are more regular, and which are
better candidates for being written in tengwar (I'd nominate Nivkh,
or any of a number of Australian languages such as Aranta).
But no primary-order tengwa for {'}, please! Writing it as if
it were a voiceless velar fricative would make it more similar
to {k} or {x} that many other consonants would be. If it must
be a tengwa, let it be a very distinctive one such as yanta.
(This is why I've always opposed {h} for {'} in Roman script
-- it should be very prominent, but it is a Good Thing that it
doesn't look like a consonant.)
And if you want to take advantage of the logicality of the tengwar,
you shouldn't dream of writing {x} as if it were a voiced consonant.
--
"mu' Dajatlhpa', reH DajatlhlaH, <soxan tA nagoftI, tawAnI-^s goft,
'ach Dajatlhpu'DI', DughatlhlaH" walI gofteH rA bAz natwAn nehoft>
(Sheikh Muslihuddin Abu Muhammad Abdullah Saadi Shirazi)
Ivan A Derzhanski <http://www.math.bas.bg/~iad/>
H: cplx Iztok bl 91, 1113 Sofia, Bulgaria <i...@math.bas.bg>
W: Dept for Math Lx, Inst for Maths & CompSci, Bulg Acad of Sciences
Ivan A Derzhanski
> Ivan A Derzhanski writes:
> > The Chinese script is always monospace, and yet looks gorgeous,
> > doesn't it?
>
> knows what they are doing. However, keep in mind that the parts of
> Chinese orthography closest to what might be called "letters" are
> *not* monospace - many Chinese characters are composed of several
> other characters, squished and cropped to fit into the "monospace"
> rectangle of a Chinese "character".
True, but still in the written form of the Chinese language there is
*something* that is a sort of unit -- albeit not a minimal one --
in three ways: (1) phonetically (a syllable), (2) morphologically (a
morpheme) and (3) spatially (a fixed width). It doesn't matter
so much that characters may or may not contain several components,
just as syllables may contain 1 to 4 phonemes (plus tone).
I'm not advocating transferring the Chinese model to just any other
language, just pointing out that spatial uniformity can work well
when it is linked to uniformity in other, language-specific aspects.
Korean groups (and squishes, in the process) letters into written
syllables, which are all of the same size. That works very well
for a language with such a well-defined syllable structure, but
I wouldn't recommend it for English or Russian.
> I think I haven't made clear one of my concerns about monospace
> fonts for lojban. It's a matter of "visual tone." [...]
> I would suggest that Chinese orthography generally exhibits
> this characteristic. I can't think of a single natural orthography
> that *doesn't* particularly exhibit this trait.
There are some, but they're not among the most widespread ones.
In books on writing systems I've certainly seen samples of some
particularly incoherent ones (usually, as in the case of Japanese,
resulting from the mix of two or more very different systems).
> Now that I think of it, though, studies on font readability (as
> I recall), indicate that what you find easiest to read is what you
> learn to read on. Hence, most Europeans find sans serif just dandy
> for book fonts, while most Americans find serif easier.
This European finds sans serif hideous as a book fount. As for
monospace, it gets much easier to read if word space needn't be
of constant width and the lines are justified.
> > Do you like the fact that all Lojban gismu are of equal length,
> > measured in phonemes/letters (5 in all, 3 consonants and 2 vowels)?
>
> Hmmm. I'm neutral on it.
I like it very much.
> > Well, I want the next logical thing: I also want them to be of
> > equal width on paper.
>
> Why is *width* the next logical thing? Why not height?
Because it is the width of the whole that is derived from the number
and the kind of the components, which is just what is uniform within
a syntactic category in Lojban.
> > I'd like every consonant to be as wide as every other, and every
> > vowel as wide as every other (perhaps less wide than the consonants).
[...]
> > It makes sense for {' , .} to be much narrower than the letters.
>
> Which nukes the uniform width bit......
It doesn't quite nuke it. It does away with the fact that, say,
{l} and {m} are of the same category and {i} is of a different one,
but {l} and {i} are as wide as one another, and much narrower than {m}.
> This is quite a stimulating discussion - I am enjoying it immensely
> (and if I sound strident or shrill anywhere, please don't take it as a
> personal attack - I don't intend such and am merely getting excited by
> an enjoyable discussion).
My position exactly.
David Brookshire Conner
Ivan A Derzhanski writes:
> From: Ivan A Derzhanski <i...@math.bas.bg>
>
> Yes, I am on the list, and here's what I have to say:
>
> (1) Like Mark, I prefer monospace to proportional for Lojban.
> I don't know whether Lojban is hot or cold, poetic or prosaic,
> fluid or solid, nor whether it needs to be any of those things.
> Its first and strongest claim is to being a *logical* language,
> and in this it can't afford to fail.
Point taken.
> Forget about typewritten English; think on a larger scale.
> The Chinese script is always monospace, and yet looks gorgeous,
> doesn't it?
Yes, it does look gorgeous, especially when written by someone that
knows what they are doing. However, keep in mind that the parts of
Chinese orthography closest to what might be called "letters" are
*not* monospace - many Chinese characters are composed of several
other characters, squished and cropped to fit into the "monospace"
rectangle of a Chinese "character".
I think I haven't made clear one of my concerns about monospace fonts
for lojban. It's a matter of "visual tone." If you recall Knuth's
writing on the line breaking algorithm for TeX, that's the same kind
of thing I mean - the main prose should strive for a relatively even
"color". From what I've read on typography, this is a major concern
when designing a book font (for display fonts, all bets are off, of
course). I would suggest that Chinese orthography generally exhibits
this characteristic. I can't think of a single natural orthography
that *doesn't* particularly exhibit this trait. The only one that
comes to mind (though this may just be my lack of truly broad
familiarity with the field) is Japanese, which only lacks a uniform
"color" because written Japanese is usually mix of *four* orthographic
systems (one matching Chinese orthography, two Japanese syllabic
orthographies (one for native words and one for foreign words), and
the Roman alphabet).
Now that I think of it, though, studies on font readability (as I
recall), indicate that what you find easiest to read is what you learn
to read on. Hence, most Europeans find sans serif just dandy for book
fonts, while most Americans find serif easier. Perhaps this suggests
that people find monospace fonts harder to read because they generally
did not grow up learning to read that way.
> More to the point, it is a logical thing: 1 unit
> of width = 1 syllable = 1 morpheme. (With very few exceptions.)
I lost your referent here. Are you referring to Chinese still or to
lojban? If Chinese, it's incorrect, as previously noted, many of the
characters are a combination of "primitive" characters. If lojban,
yes, within a particular part of speech, if you use a monospace font
(hence this thread) or if you measure width in characters.
> Do you like the fact that all Lojban gismu are of equal length,
> measured in phonemes/letters (5 in all, 3 consonants and 2 vowels)?
Hmmm. I'm neutral on it. It's a nice pattern, especially when the
letters are chosen cleverly.
> Well, I want the next logical thing: I also want them to be of
> equal width on paper.
Why is *width* the next logical thing? Why not height? Or color - you
could track cmavo tighter than brivla or some other pattern.
> That doesn't have to mean monospace as in
> all characters being of the same width, but I'd like every consonant
> to be as wide as every other, and every vowel as wide as every other
> (perhaps less wide than the consonants).
If logic becomes a concern for the orthography, though, it seems to
open things up to some sticky questions - the lojban alphabet seems to
be based only scantly on "logic" - its basis is much stronger in
engineering and pragmatics. We use it because many people already use
it.
Hmmm. This has the gears going. What would a lojban-specific,
*logical* orthography look like?
> It makes sense for {' , .}
> to be much narrower than the letters.
Which nukes the uniform width bit......
This is quite a stimulating discussion - I am enjoying it immensely
(and if I sound strident or shrill anywhere, please don't take it as a
personal attack - I don't intend such and am merely getting excited by
an enjoyable discussion).
Brook
---------
E Pluribus Modem